Statistics and Computing

, Volume 27, Issue 3, pp 833–844

Investigation of the widely applicable Bayesian information criterion

  • N. Friel
  • J. P. McKeone
  • C. J. Oates
  • A. N. Pettitt
Article

DOI: 10.1007/s11222-016-9657-y

Cite this article as:
Friel, N., McKeone, J.P., Oates, C.J. et al. Stat Comput (2017) 27: 833. doi:10.1007/s11222-016-9657-y

Abstract

The widely applicable Bayesian information criterion (WBIC) is a simple and fast approximation to the model evidence that has received little practical consideration. WBIC uses the fact that the log evidence can be written as an expectation, with respect to a powered posterior proportional to the likelihood raised to a power \(t^*\in {(0,1)}\), of the log deviance. Finding this temperature value \(t^*\) is generally an intractable problem. We find that for a particular tractable statistical model that the mean squared error of an optimally-tuned version of WBIC with correct temperature \(t^*\) is lower than an optimally-tuned version of thermodynamic integration (power posteriors). However in practice WBIC uses the a canonical choice of \(t=1/\log (n)\). Here we investigate the performance of WBIC in practice, for a range of statistical models, both regular models and singular models such as latent variable models or those with a hierarchical structure for which BIC cannot provide an adequate solution. Our findings are that, generally WBIC performs adequately when one uses informative priors, but it can systematically overestimate the evidence, particularly for small sample sizes.

Keywords

Marginal likelihood Evidence Power posteriors Widely applicable Bayesian information criterion 

Copyright information

© Springer Science+Business Media New York 2016

Authors and Affiliations

  • N. Friel
    • 1
  • J. P. McKeone
    • 3
    • 4
  • C. J. Oates
    • 2
    • 4
  • A. N. Pettitt
    • 3
    • 4
  1. 1.School of Mathematics and Statistics and Insight Centre for Data AnalyticsUniversity College DublinDublinIreland
  2. 2.School of Mathematical and Physical SciencesUniversity of Technology SydneySydneyAustralia
  3. 3.School of Mathematical SciencesQueensland University of TechnologyBrisbaneAustralia
  4. 4.Australian Research Council Centre for Excellence in Mathematical and Statistical FrontiersParkvilleAustralia